Mercurial > hg > mpdl-group

comparison software/eXist/webapp/mpdl/_stuff/oxygen-projects/monte-project/page-query-result.xquery.old.xquery @ 7:5589d865af7a

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Erstellung XQL/XSL Applikation
author Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
date Tue, 08 Feb 2011 15:16:46 +0100
parents
children
comparison
equal deleted inserted replaced
6:2396a569e446 7:5589d865af7a
1 xquery version "1.0";
2
3 module namespace ft="http://exist-db.org/xquery/lucene";
4 declare namespace request="http://exist-db.org/xquery/request";
5 declare namespace text="http://exist-db.org/xquery/text";
6 declare namespace transform="http://exist-db.org/xquery/transform";
7 declare namespace util="http://exist-db.org/xquery/util";
8 declare namespace docc = "http://www.mpiwg-berlin.mpg.de/ns/mpdl/docc";
9 declare namespace time-util = "http://www.mpiwg-berlin.mpg.de/ns/mpdl/time-util";
10
11 declare function time-util:duration-as-ms($t) {
12 round((minutes-from-duration($t) * 60 + seconds-from-duration($t)) * 1000 )
13 };
14
15 declare function docc:ft-search($docPath, $ftQuery) as node() {
16 let $document := doc($docPath)
17 (: performance reasons: all hits (not only the first 10! ) are passed through the :)
18 (: for loop: so the overhead in each loop has to be minimized :)
19 let $t := $document//s[ft:query(., $ftQuery)]
20 let $tempQueryResult :=
21 for $ss at $poss in $t
22 (: where $poss > 100 and $poss <= 200 :)
23 return $ss
24 let $queryResult :=
25 for $s at $pos in $tempQueryResult
26 let $pnOfS := count($document//pb[. << $s]) (: faster: comparison only in pb elements of this document :)
27 let $posOfS := count($document//pb[$pnOfS]/following::s[. << $s]) + 1 (: faster: comparisonon only in s elements of this document :)
28 let $resultElem :=
29 <hit>
30 <pos>{$pos}</pos>
31 <pn>{$pnOfS}</pn>
32 <pos-of-s>{$posOfS}</pos-of-s>
33 {$s}
34 </hit>
35 return $resultElem
36 let $resultSize := count($queryResult)
37 let $result :=
38 <ft-query>
39 <name>{$ftQuery}</name>
40 <result>
41 <size>{$resultSize}</size>
42 <hits>
43 {$queryResult}
44 </hits>
45 </result>
46 </ft-query>
47 return $result
48 };
49
50 declare function docc:termEntries($term as xs:string, $data as xs:int+) {
51 let $result :=
52 <entry>
53 <term>{$term}</term>
54 <frequency>{$data[1]}</frequency>
55 <documents>{$data[2]}</documents>
56 <position>{$data[3]}</position>
57 <rank>{$data[4]}</rank>
58 </entry>
59 return $result
60 };
61
62 declare function docc:indexTerms($document, $indexTermsStartStr, $count) as node()* {
63 let $docName := util:document-name($document)
64 let $index := /archimedes/info/cvs_file[contains(., $docName)]/root()/archimedes/text
65 let $callback := util:function(QName("http://www.mpiwg-berlin.mpg.de/ns/mpdl/docc", "docc:termEntries"), 2)
66 let $result := text:index-terms($index, $indexTermsStartStr, $callback, $count)
67 return
68 <index-terms>
69 <start-string>{$indexTermsStartStr}</start-string>
70 {$result}
71 </index-terms>
72 };
73
74
75 let $document := request:get-parameter("document", "agric_remet_001_la_1556")
76 let $mode := request:get-parameter("mode", "image")
77 let $pn := number(request:get-parameter("pn", "1"))
78 let $ftQuery := request:get-parameter("ft-query", "")
79 let $indexTermsStartStr := request:get-parameter("index-terms-start-str", "")
80
81 let $currentTimeBegin := util:system-time()
82 let $docPath := concat("/db/archimedes/", $document, ".xml")
83 let $docRoot := doc($docPath)
84 let $archimedesInfo := $docRoot/archimedes/info
85 let $locator := $archimedesInfo/locator
86 let $countPages := count($docRoot//pb)
87 let $pb1 := $docRoot//pb[$pn]
88 let $positionOfFirstFigureAfterPB1 := count($pb1/following::figure[1]/preceding::figure) + 1 (: position der ersten figure nach pb1 :)
89 let $pb1ImageRefLeft := substring-before($locator, ".")
90 let $pb1ImageRefRight := "01"
91 let $pb2 := $docRoot//pb[$pn + 1]
92 let $errorCode :=
93 if ($pn < $countPages + 1 and $pn > 0)
94 then 0
95 else 1
96
97 let $retPageFragment :=
98 if ($mode = "image" or $errorCode > 0)
99 then ()
100 else util:get-fragment-between($pb1, $pb2, true())
101
102 let $returnPageFragment := util:parse($retPageFragment) (: string2xml: returns a valid xml document for that string :)
103
104 let $ftHits :=
105 if ($ftQuery = "")
106 then ()
107 else docc:ft-search($docPath, $ftQuery)
108
109 let $indexTerms :=
110 if($indexTermsStartStr = "")
111 then ()
112 else docc:indexTerms($docRoot, $indexTermsStartStr, 100)
113
114 let $currentTimeEnd := util:system-time()
115 let $neededTime := time-util:duration-as-ms($currentTimeEnd - $currentTimeBegin)
116
117 let $xmlResult :=
118 if ($errorCode = 0)
119 then
120 <result>
121 <mode>{$mode}</mode>
122 <document-description>
123 <document-name>{$document}</document-name>
124 {$archimedesInfo}
125 <count-pages>{$countPages}</count-pages>
126 <performance>{$neededTime}</performance>
127 </document-description>
128 <page>
129 <number>{$pn}</number>
130 <image>http://nausikaa2.rz-berlin.mpg.de/digitallibrary/digilib.jsp?fn=/permanent/archimedes/{$document}/{$pb1ImageRefLeft}-{$pb1ImageRefRight}-pageimg&amp;pn={$pn}</image>
131 <image-echo>http://echo.mpiwg-berlin.mpg.de/zogilib?fn=/permanent/archimedes/{$document}/{$pb1ImageRefLeft}-{$pb1ImageRefRight}-pageimg&amp;pn={$pn}</image-echo>
132 <image-scaler>http://nausikaa2.rz-berlin.mpg.de/digitallibrary/servlet/Scaler?fn=/permanent/archimedes/{$document}/{$pb1ImageRefLeft}-{$pb1ImageRefRight}-pageimg&amp;pn={$pn}</image-scaler>
133 <xml-url>?document={$document}&amp;pn={$pn}&amp;mode=xml</xml-url>
134 <firstFigurePosition>{$positionOfFirstFigureAfterPB1}</firstFigurePosition>
135 <content>{$returnPageFragment}</content>
136 </page>
137 {$ftHits}
138 {$indexTerms}
139 </result>
140 else if ($errorCode = 1)
141 then <error>No result: Page {$pn} not found</error>
142 else <error>undefined error: {$errorCode}</error>
143
144 let $declare :=
145 if ($mode = "text" or $mode = "image" or $mode = "xml")
146 then util:declare-option("exist:serialize", "method=html media-type=text/html omit-xml-declaration=no indent=yes encoding=UTF-8")
147 else util:declare-option("exist:serialize", "method=xml media-type=text/xml omit-xml-declaration=no indent=yes encoding=UTF-8")
148
149 let $result := transform:transform($xmlResult, doc("/db/xsl/archimedes-page-fragment.xsl"), ())
150
151 return $result